Research on Subject Knowledge Flow Based on"Problem-Method"Knowledge Units Mining
[Purpose/Significance]Research problems and research methods constitute the main characteristics of a subject.Knowledge unit is the smallest elements of knowledge content.This paper mines the knowledge units that express research problems and research methods in academic literature.It is helpful to describe the fine-grained knowledge flow characteristics of the subject and promote scientific innovation at the same time.[Method/Process]Through the automatic extraction of problem knowledge units and method knowledge units from academic litera-ture,it identifies the triplet relationship among knowledge units by a comprehensive combination of deep learning algorithm,word vector and word co-occurrence combination,hierarchical clustering algorithm,spatial distance calculation and other methods,so as to build the fine-grained knowledge flow network within the subject and sum-marize its characteristics.[Result/Conclusion]From the data of Chinese core journals in library and information in CNKI,it identifies the problem knowledge units,method knowledge units and their relationship from the abstract of the paper,and constructs the"problem-method"knowledge flow network.By this empirical analysis,it verifies the feasibility and effectiveness of the model.
NETL
NSTL
万方数据
